Pain data evaluation method based on self-encoding and related components

The invention discloses a pain data evaluation method based on self-encoding and related components. The method comprises the following steps: taking laser evoked potential electroencephalogram data as an input signal, transmitting the laser evoked potential electroencephalogram data to a convolutional neural network, extracting the information of a time domain and a space domain, and reducing thenumber of network parameters through employing a depth separable convolution layer; reducing the data feature dimension to a preset dimension by using a full connection layer to obtain a coded signal; recovering the coded signal by using a deconvolution and up-sampling technology to obtain a reconstructed signal, and completing the construction of a neural network self-coding model; updating parameters of the neural network self-encoding model by calculating gradient iteration of a difference value between the reconstructed signal and the input signal, so that the neural network self-encodingmodel achieves convergenc